In alignment-inducing multi-agent settings, LLM agents show decision divergence between public and off-the-record channels rising from a 3% baseline to roughly 40%, consistent across stance, semantic, NLI, and survey measures.
Two Tales of Persona in LLM s: A Survey of Role-Playing and Personalization
13 Pith papers cite this work. Polarity classification is still indexing.
13
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
roles
background 1polarities
background 1representative citing papers
LegalWorld is a life-cycle interactive environment modeling Chinese civil litigation as five causally connected stages grounded in 75,309 judgments, paired with LongJud-Bench for cross-stage agent evaluation.
ZIPP conditions diffusion models on LLM-rewritten prompts derived from graph-mined natural-language personas to achieve zero-shot personalization, reporting 13-20% gains and 79% human preference win rate over generic outputs.
CoPersona introduces a multiplex persona graph for facet-level peer alignment and a dual-branch retrieval-plus-reasoning architecture to improve LLM personalization under sparse and biased user interaction data.
Persona-driven generations by LLMs in MCQA tasks exhibit instability that differs systematically by model family, size, domain, and prompt format.
Psy-CoT decomposes reasoning into Interaction Perception, Psychological Empathy, and Logical Construction while RAPO asymmetrically weights role-specific tokens during policy optimization, outperforming prior CoT and GRPO baselines on role-playing benchmarks.
REVERIEMEM is a three-layer perspective-bounded memory system that raises knowledge boundary fidelity by 34.6 points and wins ~79% of narrative comparisons on a new book-based role-playing benchmark.
IVIE generates complete playable interactive fiction worlds via a four-stage incremental pipeline that combines LLM creativity with symbolic validation for coherence.
GenPT applies generative projective testing to LLM agents and reports lower directional bias plus greater longitudinal sensitivity than self-report questionnaires.
Profile-conditioned LLMs achieve higher tacit alignment with humans on subjective spectra when traits match, as quantified by the new Tacit Understanding Index (TUX) from 241 humans and 200 agents.
An extended annotation scheme with new categories and attributes plus a Gemma-300M-based multi-head classifier achieves 81.6% macro F1 on personal fact classification, outperforming few-shot LLM baselines by nearly 9 points with lower compute.
TSUBASA improves long-horizon personalization in LLMs via dynamic memory evolution for writing and context-distillation self-learning for reading, outperforming Mem0 and Memory-R1 on Qwen-3 benchmarks while reducing token use.
Synthia creates scalable personas from Bluesky posts that better match human survey responses than prior methods, uses smaller models, and retains social network structure for network-aware analysis.
citing papers explorer
-
What LLM Agents Say When No One Is Watching: Social Structure and Latent Objective Emergence in Multi-Agent Debates
In alignment-inducing multi-agent settings, LLM agents show decision divergence between public and off-the-record channels rising from a 3% baseline to roughly 40%, consistent across stance, semantic, NLI, and survey measures.
-
LegalWorld: A Life-Cycle Interactive Environment for Legal Agents
LegalWorld is a life-cycle interactive environment modeling Chinese civil litigation as five causally connected stages grounded in 75,309 judgments, paired with LongJud-Bench for cross-stage agent evaluation.
-
ZIPP:Zero-shot Image Personalization from Personas
ZIPP conditions diffusion models on LLM-rewritten prompts derived from graph-mined natural-language personas to achieve zero-shot personalization, reporting 13-20% gains and 79% human preference win rate over generic outputs.
-
CoPersona: Collaborative Persona Graphs for Robust LLM Personalization
CoPersona introduces a multiplex persona graph for facet-level peer alignment and a dual-branch retrieval-plus-reasoning architecture to improve LLM personalization under sparse and biased user interaction data.
-
Persona Non Grata: LLM Persona-Driven Generations in MCQA are Unstable in Distinct Dimensions
Persona-driven generations by LLMs in MCQA tasks exhibit instability that differs systematically by model family, size, domain, and prompt format.
-
Improving General Role-Playing Agents via Psychology-Grounded Reasoning and Role-Aware Policy Optimization
Psy-CoT decomposes reasoning into Interaction Perception, Psychological Empathy, and Logical Construction while RAPO asymmetrically weights role-specific tokens during policy optimization, outperforming prior CoT and GRPO baselines on role-playing benchmarks.
-
Staying In Character: Perspective-Bounded Memory For Book-Based Role-Playing Agents
REVERIEMEM is a three-layer perspective-bounded memory system that raises knowledge boundary fidelity by 34.6 points and wins ~79% of narrative comparisons on a new book-based role-playing benchmark.
-
IVIE: A Neuro-symbolic Approach to Incremental and Validated Generation of Interactive Fiction Worlds
IVIE generates complete playable interactive fiction worlds via a four-stage incremental pipeline that combines LLM creativity with symbolic validation for coherence.
-
GenPT: Beyond Self-Report for Reliable LLM Psychometrics via Generative Projective Testing
GenPT applies generative projective testing to LLM agents and reports lower directional bias plus greater longitudinal sensitivity than self-report questionnaires.
-
TUX: Measuring Human--AI Tacit Understanding
Profile-conditioned LLMs achieve higher tacit alignment with humans on subjective spectra when traits match, as quantified by the new Tacit Understanding Index (TUX) from 241 humans and 200 agents.
-
TSUBASA: Improving Long-Horizon Personalization via Evolving Memory and Self-Learning with Context Distillation
TSUBASA improves long-horizon personalization in LLMs via dynamic memory evolution for writing and context-distillation self-learning for reading, outperforming Mem0 and Memory-R1 on Qwen-3 benchmarks while reducing token use.
-
Synthia: Scalable Grounded Persona Generation from Social Media Data
Synthia creates scalable personas from Bluesky posts that better match human survey responses than prior methods, uses smaller models, and retains social network structure for network-aware analysis.